This adds a BPF SK_MSG program helper so that we can pop data from a
msg. We use this to pop metadata from a previous push data call.
Signed-off-by: NJohn Fastabend <john.fastabend@gmail.com>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

'offset' is constant and if it is zero, no need to subtract it
from BPF_REG_TMP.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NDaniel Borkmann <daniel@iogearbox.net>

I do not see how one can effectively use skb_insert() without holding
some kind of lock. Otherwise other cpus could have changed the list
right before we have a chance of acquiring list->lock.

Only existing user is in drivers/infiniband/hw/nes/nes_mgt.c and this
one probably meant to use __skb_insert() since it appears nesqp->pau_list
is protected by nesqp->pau_lock. This looks like nesqp->pau_lock
could be removed, since nesqp->pau_list.lock could be used instead.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Cc: Faisal Latif <faisal.latif@intel.com>
Cc: Doug Ledford <dledford@redhat.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: linux-rdma <linux-rdma@vger.kernel.org>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

A recent change added a null check on p->dev after p->dev was being
dereferenced by the ns_capable check on p->dev. It turns out that
neither the p->dev and p->br null checks are necessary, and can be
removed, which cleans up a static analyis warning.

As Nikolay Aleksandrov noted, these checks can be removed because:

"My reasoning of why it shouldn't be possible:
- On port add new_nbp() sets both p->dev and p->br before creating
  kobj/sysfs

- On port del (trickier) del_nbp() calls kobject_del() before call_rcu()
  to destroy the port which in turn calls sysfs_remove_dir() which uses
  kernfs_remove() which deactivates (shouldn't be able to open new
  files) and calls kernfs_drain() to drain current open/mmaped files in
  the respective dir before continuing, thus making it impossible to
  open a bridge port sysfs file with p->dev and p->br equal to NULL.

So I think it's safe to remove those checks altogether. It'd be nice to
get a second look over my reasoning as I might be missing something in
sysfs/kernfs call path."

Thanks to Nikolay Aleksandrov's suggestion to remove the check and
David Miller for sanity checking this.

Detected by CoverityScan, CID#751490 ("Dereference before null check")

Fixes: a5f3ea54 ("net: bridge: add support for raw sysfs port options")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Acked-by: NNikolay Aleksandrov <nikolay@cumulusnetworks.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Due to an explicit check in rocker_world_port_obj_vlan_add(),
dsa_slave_switchdev_event() resp. port_switchdev_event(), VLAN objects
that are added to a device that is not a front-panel port device are
ignored. Therefore this check is immaterial.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Drop switchdev_ops.switchdev_port_obj_add and _del. Drop the uses of
this field from all clients, which were migrated to use switchdev
notification in the previous patches.

Add a new function switchdev_port_obj_notify() that sends the switchdev
notifications SWITCHDEV_PORT_OBJ_ADD and _DEL.

Update switchdev_port_obj_del_now() to dispatch to this new function.
Drop __switchdev_port_obj_add() and update switchdev_port_obj_add()
likewise.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After the transition from switchdev operations to notifier chain (which
will take place in following patches), the onus is on the driver to find
its own devices below possible layer of LAG or other uppers.

The logic to do so is fairly repetitive: each driver is looking for its
own devices among the lowers of the notified device. For those that it
finds, it calls a handler. To indicate that the event was handled,
struct switchdev_notifier_port_obj_info.handled is set. The differences
lie only in what constitutes an "own" device and what handler to call.

Therefore abstract this logic into two helpers,
switchdev_handle_port_obj_add() and switchdev_handle_port_obj_del(). If
a driver only supports physical ports under a bridge device, it will
simply avoid this layer of indirection.

One area where this helper diverges from the current switchdev behavior
is the case of mixed lowers, some of which are switchdev ports and some
of which are not. Previously, such scenario would fail with -EOPNOTSUPP.
The helper could do that for lowers for which the passed-in predicate
doesn't hold. That would however break the case that switchdev ports
from several different drivers are stashed under one master, a scenario
that switchdev currently happily supports. Therefore tolerate any and
all unknown netdevices, whether they are backed by a switchdev driver
or not.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Following patches will change the way of distributing port object
changes from a switchdev operation to a switchdev notifier. The
switchdev code currently recursively descends through layers of lower
devices, eventually calling the op on a front-panel port device. The
notifier will instead be sent referencing the bridge port device, which
may be a stacking device that's one of front-panel ports uppers, or a
completely unrelated device.

DSA currently doesn't support any other uppers than bridge.
SWITCHDEV_OBJ_ID_HOST_MDB and _PORT_MDB objects are always notified on
the bridge port device. Thus the only case that a stacked device could
be validly referenced by port object notifications are bridge
notifications for VLAN objects added to the bridge itself. But the
driver explicitly rejects such notifications in dsa_port_vlan_add(). It
is therefore safe to assume that the only interesting case is that the
notification is on a front-panel port netdevice. Therefore keep the
filtering by dsa_slave_dev_check() in place.

To handle SWITCHDEV_PORT_OBJ_ADD and _DEL, subscribe to the blocking
notifier chain. Dispatch to rocker_port_obj_add() resp. _del() to
maintain the behavior that the switchdev operation based code currently
has.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In general one can't assume that a switchdev notifier is called in a
non-atomic context, and correspondingly, the switchdev notifier chain is
an atomic one.

However, port object addition and deletion messages are delivered from a
process context. Even the MDB addition messages, whose delivery is
scheduled from atomic context, are queued and the delivery itself takes
place in blocking context. For VLAN messages in particular, keeping the
blocking nature is important for error reporting.

Therefore introduce a blocking notifier chain and related service
functions to distribute the notifications for which a blocking context
can be assumed.
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Reviewed-by: NJiri Pirko <jiri@mellanox.com>
Reviewed-by: NIdo Schimmel <idosch@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When an rmb is no longer in use by a connection, unregister its rkey at
the remote peer with an LLC DELETE RKEY message. With this change,
unused buffers held in the buffer pool are no longer registered at the
remote peer. They are registered before the buffer is actually used and
unregistered when they are no longer used by a connection.
Signed-off-by: NKarsten Graul <kgraul@linux.ibm.com>
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Add the infrastructure to send LLC messages of type DELETE RKEY to
unregister a shared memory region at the peer.
Signed-off-by: NKarsten Graul <kgraul@linux.ibm.com>
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a send failed then don't start to wait for a response in
smc_llc_do_confirm_rkey.
Signed-off-by: NKarsten Graul <kgraul@linux.ibm.com>
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

For easier reading move the unlock of mutex smc_create_lgr_pending into
smc_listen_work(), i.e. into the function the mutex has been locked.
No functional change.
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

After sending one of the initial LLC messages CONFIRM LINK or
ADD LINK, there is already a wait for the LLC response. It does
not make sense to wait another long time for a CLC DECLINE. Thus
this patch introduces a shorter wait time for these cases.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If a link is terminated that has never reached the active state,
there is no need to trigger an LLC DELETE LINK.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

If connection initialization fails for the LLC CONFIRM LINK or the
LLC ADD LINK step, fallback to TCP should be enabled. Thus
the negative return code -EAGAIN should switch to a positive timeout
reason code in these cases, and the internal CLC socket should
not have a set sk_err.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There is no need to store the return value in sk_err, if it is
afterwards cleared again with sock_error(). This patch sets the
return value directly. Just cleanup, no functional change.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

smc_lgr_free() is just called inside smc_core.c. Make it static.
Just cleanup, no functional change.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The tcp_listen_worker is already initialized when socket is
created (in smc_sock_alloc()). Get rid of the duplicate
initialization in smc_listen(). No functional change.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

We very often have few flows/chains to look at, and we
might increase GRO_HASH_BUCKETS to 32 or 64 in the future.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

commit f2cbd485 ("net/sched: act_police: fix race condition on state
variables") introduces a new spinlock, but forgets its initialization.
Ensure that tcf_police_init() initializes 'tcfp_lock' every time a 'police'
action is newly created, to avoid the following lockdep splat:

 INFO: trying to register non-static key.
 the code is fine but needs lockdep annotation.
 turning off the locking correctness validator.
 <...>
 Call Trace:
  dump_stack+0x85/0xcb
  register_lock_class+0x581/0x590
  __lock_acquire+0xd4/0x1330
  ? tcf_police_init+0x2fa/0x650 [act_police]
  ? lock_acquire+0x9e/0x1a0
  lock_acquire+0x9e/0x1a0
  ? tcf_police_init+0x2fa/0x650 [act_police]
  ? tcf_police_init+0x55a/0x650 [act_police]
  _raw_spin_lock_bh+0x34/0x40
  ? tcf_police_init+0x2fa/0x650 [act_police]
  tcf_police_init+0x2fa/0x650 [act_police]
  tcf_action_init_1+0x384/0x4c0
  tcf_action_init+0xf6/0x160
  tcf_action_add+0x73/0x170
  tc_ctl_action+0x122/0x160
  rtnetlink_rcv_msg+0x2a4/0x490
  ? netlink_deliver_tap+0x99/0x400
  ? validate_linkmsg+0x370/0x370
  netlink_rcv_skb+0x4d/0x130
  netlink_unicast+0x196/0x230
  netlink_sendmsg+0x2e5/0x3e0
  sock_sendmsg+0x36/0x40
  ___sys_sendmsg+0x280/0x2f0
  ? _raw_spin_unlock+0x24/0x30
  ? handle_pte_fault+0xafe/0xf30
  ? find_held_lock+0x2d/0x90
  ? syscall_trace_enter+0x1df/0x360
  ? __sys_sendmsg+0x5e/0xa0
  __sys_sendmsg+0x5e/0xa0
  do_syscall_64+0x60/0x210
  entry_SYSCALL_64_after_hwframe+0x49/0xbe
 RIP: 0033:0x7f1841c7cf10
 Code: c3 48 8b 05 82 6f 2c 00 f7 db 64 89 18 48 83 cb ff eb dd 0f 1f 80 00 00 00 00 83 3d 8d d0 2c 00 00 75 10 b8 2e 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 31 c3 48 83 ec 08 e8 ae cc 00 00 48 89 04 24
 RSP: 002b:00007ffcf9df4d68 EFLAGS: 00000246 ORIG_RAX: 000000000000002e
 RAX: ffffffffffffffda RBX: 0000000000000001 RCX: 00007f1841c7cf10
 RDX: 0000000000000000 RSI: 00007ffcf9df4dc0 RDI: 0000000000000003
 RBP: 000000005bf56105 R08: 0000000000000002 R09: 00007ffcf9df8edc
 R10: 00007ffcf9df47e0 R11: 0000000000000246 R12: 0000000000671be0
 R13: 00007ffcf9df4e84 R14: 0000000000000008 R15: 0000000000000000

Fixes: f2cbd485 ("net/sched: act_police: fix race condition on state variables")
Reported-by: NCong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: NDavide Caratti <dcaratti@redhat.com>
Acked-by: NCong Wang <xiyou.wangcong@gmail.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Eric noted that with UDP GRO and NAPI timeout, we could keep a single
UDP packet inside the GRO hash forever, if the related NAPI instance
calls napi_gro_complete() at an higher frequency than the NAPI timeout.
Willem noted that even TCP packets could be trapped there, till the
next retransmission.
This patch tries to address the issue, flushing the old packets -
those with a NAPI_GRO_CB age before the current jiffy - before scheduling
the NAPI timeout. The rationale is that such a timeout should be
well below a jiffy and we are not flushing packets eligible for sane GRO.

v1  -> v2:
 - clarified the commit message and comment

RFC -> v1:
 - added 'Fixes tags', cleaned-up the wording.
Reported-by: NEric Dumazet <eric.dumazet@gmail.com>
Fixes: 3b47d303 ("net: gro: add a per device gro flush timer")
Signed-off-by: NPaolo Abeni <pabeni@redhat.com>
Acked-by: NWillem de Bruijn <willemb@google.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When we add a new IPv6 address, we should also join corresponding solicited-node
multicast address, unless the interface has IFF_NOARP flag, as function
addrconf_join_solict() did. But if we remove IFF_NOARP flag later, we do
not do dad and add the mcast address. So we will drop corresponding neighbour
discovery message that came from other nodes.

A typical example is after creating a ipvlan with mode l3, setting up an ipv6
address and changing the mode to l2. Then we will not be able to ping this
address as the interface doesn't join related solicited-node mcast address.

Fix it by re-doing dad when interface changed IFF_NOARP flag. Then we will add
corresponding mcast group and check if there is a duplicate address on the
network.
Reported-by: NJianlin Shi <jishi@redhat.com>
Reviewed-by: NStefano Brivio <sbrivio@redhat.com>
Signed-off-by: NHangbin Liu <liuhangbin@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

tpacket_snd sends packets with user pages linked into skb frags. It
notifies that pages can be reused when the skb is released by setting
skb->destructor to tpacket_destruct_skb.

This can cause data corruption if the skb is orphaned (e.g., on
transmit through veth) or cloned (e.g., on mirror to another psock).

Create a kernel-private copy of data in these cases, same as tun/tap
zerocopy transmission. Reuse that infrastructure: mark the skb as
SKBTX_ZEROCOPY_FRAG, which will trigger copy in skb_orphan_frags(_rx).

Unlike other zerocopy packets, do not set shinfo destructor_arg to
struct ubuf_info. tpacket_destruct_skb already uses that ptr to notify
when the original skb is released and a timestamp is recorded. Do not
change this timestamp behavior. The ubuf_info->callback is not needed
anyway, as no zerocopy notification is expected.

Mark destructor_arg as not-a-uarg by setting the lower bit to 1. The
resulting value is not a valid ubuf_info pointer, nor a valid
tpacket_snd frame address. Add skb_zcopy_.._nouarg helpers for this.

The fix relies on features introduced in commit 52267790 ("sock:
add MSG_ZEROCOPY"), so can be backported as is only to 4.14.

Tested with from `./in_netns.sh ./txring_overwrite` from
http://github.com/wdebruij/kerneltools/tests

Fixes: 69e3c75f ("net: TX_RING and packet mmap")
Reported-by: NAnand H. Krishnan <anandhkrishnan@gmail.com>
Signed-off-by: NWillem de Bruijn <willemb@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This could be used to rate limit egress traffic in concert with a qdisc
which supports Earliest Departure Time, such as FQ.

Write access from cg skb progs only with CAP_SYS_ADMIN, since the value
will be used by downstream qdiscs. It might make sense to relax this.

Changes v1 -> v2:
  - allow access from cg skb, write only with CAP_SYS_ADMIN
Signed-off-by: NVlad Dumitrescu <vladum@google.com>
Acked-by: NEric Dumazet <edumazet@google.com>
Acked-by: NWillem de Bruijn <willemb@google.com>
Signed-off-by: NAlexei Starovoitov <ast@kernel.org>

Allow querying bridge port flags so that drivers capable of performing
VxLAN learning will update the bridge driver only if learning is enabled
on its bridge port corresponding to the VxLAN device.
Signed-off-by: NIdo Schimmel <idosch@mellanox.com>
Reviewed-by: NPetr Machata <petrm@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In smc_wr_tx_put_slot() field pend->idx is used after being
cleared. That means always idx 0 is cleared in the wr_tx_mask.
This results in a broken administration of available WR send
payload buffers.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Running uperf tests with SMCD on LPARs results in corrupted cursors.
SMCD cursors should be treated atomically to fix cursor corruption.
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a SMC-D link group is freed, a shutdown signal should be sent to
the peer to indicate that the link group is invalid. This patch adds the
shutdown signal to the SMC code.
Signed-off-by: NHans Wippel <hwippel@linux.ibm.com>
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When searching for an existing link group the queue pair number is also
to be taken into consideration. When the SMC server sends a new number
in a CLC packet (keeping all other values equal) then a new link group
is to be created on the SMC client side.
Signed-off-by: NKarsten Graul <kgraul@linux.ibm.com>
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In case of a non-blocking SMC socket, the initial CLC handshake is
performed over a blocking TCP connection in a worker. If the SMC socket
is released, smc_release has to wait for the blocking CLC socket
operations (e.g., kernel_connect) inside the worker.

This patch aborts a CLC connection when the respective non-blocking SMC
socket is released to avoid waiting on socket operations or timeouts.
Signed-off-by: NHans Wippel <hwippel@linux.ibm.com>
Signed-off-by: NUrsula Braun <ubraun@linux.ibm.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Jean-Louis reported a TCP regression and bisected to recent SACK
compression.

After a loss episode (receiver not able to keep up and dropping
packets because its backlog is full), linux TCP stack is sending
a single SACK (DUPACK).

Sender waits a full RTO timer before recovering losses.

While RFC 6675 says in section 5, "Algorithm Details",

   (2) If DupAcks < DupThresh but IsLost (HighACK + 1) returns true --
       indicating at least three segments have arrived above the current
       cumulative acknowledgment point, which is taken to indicate loss
       -- go to step (4).
...
   (4) Invoke fast retransmit and enter loss recovery as follows:

there are old TCP stacks not implementing this strategy, and
still counting the dupacks before starting fast retransmit.

While these stacks probably perform poorly when receivers implement
LRO/GRO, we should be a little more gentle to them.

This patch makes sure we do not enable SACK compression unless
3 dupacks have been sent since last rcv_nxt update.

Ideally we should even rearm the timer to send one or two
more DUPACK if no more packets are coming, but that will
be work aiming for linux-4.21.

Many thanks to Jean-Louis for bisecting the issue, providing
packet captures and testing this patch.

Fixes: 5d9f4262 ("tcp: add SACK compression")
Reported-by: NJean-Louis Dupond <jean-louis@dupond.be>
Tested-by: NJean-Louis Dupond <jean-louis@dupond.be>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Acked-by: NNeal Cardwell <ncardwell@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

When a packet is trapped and the corresponding SKB marked as
already-forwarded, it retains this marking even after it is forwarded
across veth links into another bridge. There, since it ingresses the
bridge over veth, which doesn't have offload_fwd_mark, it triggers a
warning in nbp_switchdev_frame_mark().

Then nbp_switchdev_allowed_egress() decides not to allow egress from
this bridge through another veth, because the SKB is already marked, and
the mark (of 0) of course matches. Thus the packet is incorrectly
blocked.

Solve by resetting offload_fwd_mark() in skb_scrub_packet(). That
function is called from tunnels and also from veth, and thus catches the
cases where traffic is forwarded between bridges and transformed in a
way that invalidates the marking.

Fixes: 6bc506b4 ("bridge: switchdev: Add forward mark support for stacked devices")
Fixes: abf4bb6b ("skbuff: Add the offload_mr_fwd_mark field")
Signed-off-by: NPetr Machata <petrm@mellanox.com>
Suggested-by: NIdo Schimmel <idosch@mellanox.com>
Acked-by: NJiri Pirko <jiri@mellanox.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

after 'police' configuration parameters were converted to use RCU instead
of spinlock, the state variables used to compute the traffic rate (namely
'tcfp_toks', 'tcfp_ptoks' and 'tcfp_t_c') are erroneously read/updated in
the traffic path without any protection.

Use a dedicated spinlock to avoid race conditions on these variables, and
ensure proper cache-line alignment. In this way, 'police' is still faster
than what we observed when 'tcf_lock' was used in the traffic path _ i.e.
reverting commit 2d550dba ("net/sched: act_police: don't use spinlock
in the data path"). Moreover, we preserve the throughput improvement that
was obtained after 'police' started using per-cpu counters, when 'avrate'
is used instead of 'rate'.

Changes since v1 (thanks to Eric Dumazet):
- call ktime_get_ns() before acquiring the lock in the traffic path
- use a dedicated spinlock instead of tcf_lock
- improve cache-line usage

Fixes: 2d550dba ("net/sched: act_police: don't use spinlock in the data path")
Reported-and-suggested-by: NEric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: NDavide Caratti <dcaratti@redhat.com>
Reviewed-by: NEric Dumazet <edumazet@google.com>

During tcp coalescing ensure that the skb hardware timestamp refers to the
highest sequence number data.
Previously only the software timestamp was updated during coalescing.
Signed-off-by: NStephen Mallon <stephen.mallon@sydney.edu.au>
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Under stress, softirq rx handler often hits a socket owned by the user,
and has to queue the packet into socket backlog.

When this happens, skb dst refcount is taken before we escape rcu
protected region. This is done from __sk_add_backlog() calling
skb_dst_force().

Consumer will have to perform the opposite costly operation.

AFAIK nothing in tcp stack requests the dst after skb was stored
in the backlog. If this was the case, we would have had failures
already since skb_dst_force() can end up clearing skb dst anyway.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This has no value whatsoever.
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

There are two cases were we can avoid calling ktime_get_ns() :

1) Queue is empty.
2) Internal queue is not empty.
Signed-off-by: NEric Dumazet <edumazet@google.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In case of egress offloads the class/flowid assigned by the filter
may be very important for offloaded Qdisc selection.  Provide this
info to drivers.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: NJohn Hurley <john.hurley@netronome.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Allow drivers which offload GRED to report back statistics.  Since
A lot of GRED stats is fairly ad hoc in nature pass to drivers the
standard struct gnet_stats_basic/gnet_stats_queue pairs, and
untangle the values in the core.
Signed-off-by: NJakub Kicinski <jakub.kicinski@netronome.com>
Reviewed-by: NJohn Hurley <john.hurley@netronome.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>